Microsoft’s chief environmental officer on why we need a Planetary Computer
“Think of this less as a giant computer in a stark white room and more of an approach to computing that is planetary in scale and allows us to query every aspect of environmental and nature-based solutions available in real time,” Joppa wrote in a Scientific American op-ed last September. “We currently lack the data, compute power and scalability to do so. Only when we have a massive amount of planetary data and compute at a similar scale can we begin to answer one of the most complex questions ever posed—how do we manage the earth’s natural resources equitably and sustainably to ensure a prosperous and climate-stable future?”
As soon as humans could take flight in balloons and tie cameras to kites, we’ve been trying to capture and understand the world from above. The Planetary Computer takes that idea to its logical conclusion, serving as a way to envision everything happening on the planet at once with a combination of satellite imagery, sensors in the air and even robots on the ground.
Microsoft is clearly building on its AI for Earth initiative, which launched in 2018 as a way to use AI tools to help environmental organizations. Joppa says the program now covers over 500 grantees across 80 countries. “Many of those grantees have done incredible things,” he said. “But when we step back… It was really clear that we weren’t allowing our grantees or anyone else, for that matter, to build on the work of each other. To take data that was contributed by one organization, machine learning algorithms that were contributed by another, and put those two together to create a new service that anybody could use. These are all like isolated pockets of information.”
Joppa describes that realization as a humbling experience — his dream project for the last few years basically had no way to scale. So, the idea for a unified data platform was born. Getting there will take some time though. Joppa sees the Planetary Computer as something that could be powered by the data collection tools we already use, but there’s also an urgent need for new technology. Microsoft Research’s Project Premonition, for example, can help track the movement of pathogens with robots that can catch mosquitoes and other bugs, while also listening and chemically “smelling” the environment. And the move towards quantum computing, which Joppa describes as “nature’s code,” could also help us interpret the vast amounts of data that the Planetary Computer receives.
“My ultimate question… is thinking about what humanity’s objective function should be for planet Earth. What do we want to achieve with all of our natural resources?” Joppa said. “And how can we do that in a way that optimally derives human benefit from natural resources, while minimizing human negative impacts. And to do that you are actually talking about the world’s largest optimization experiment.”
There are, of course, potentially enormous issues with such a massive data collection project. Mainly, it’s worrying to think of what governments and intelligence agencies will do with it. We’ve already seen the NSA leap at the chance to invade privacy with its PRISM surveillance network. The prospect of that agency having direct access to everything happening on Earth in real-time is downright terrifying. And what about potentially hard truths? What if it recommends that we keep a nuclear reactor online, despite vocal anti-nuclear opposition, when the replacement power source is likely far more harmful to the environment? (Something New York is currently facing as it plans to close the Indian Point nuclear plant in 2021.)
“I’m prepared for getting to the outcome we want,” Joppa said when asked about challenging results from the Planetary Computer. “We just need to accept that there’s trade offs along the way. But let’s put our human ingenuity towards defining where we want to go, and what we want to be when we get there. And let’s put our computational systems that play in telling us how to most optimally travel in that direction.”